MH-1M: A 1.34 Million-Sample Multi-Feature Android Malware Dataset with Rich Metadata

Area of Science:

• Computer Science
• Cybersecurity
• Data Science

Background:

• Android malware poses a significant and evolving threat to mobile security.
• Existing datasets may lack the scale or detail required for advanced malware analysis.

Purpose of the Study:

• To introduce MH-1M, a comprehensive dataset for Android malware research.
• To facilitate advanced analysis of Android malware through open data access.

Main Methods:

• Compiled a dataset of 1,340,515 Android applications.
• Utilized the VirusTotal API with multiple detection engines for malware classification.
• Extracted features and collected extensive metadata for each application.

Main Results:

• The MH-1M dataset contains over 1.3 million Android applications.
• The dataset includes comprehensive metadata and VirusTotal analysis results.
• Over 400 GB of data is available across Figshare, Harvard Dataverse, and GitHub repositories.

Conclusions:

• The MH-1M dataset is a valuable resource for studying the current landscape of Android malware.
• Open access to this large-scale dataset will advance Android malware research and defense strategies.